Job terminated with status succses

I am runing ci for my project on self-hosted k8s runner and at first it was looks like everthing working
but suddenly I saw that test that should be failed is terminated after 15-19 minutes in the middle and all i see is “exit code 1” and the job succeeds, unfortunately I never so any logs in the k8s it look like it is a bag it the version of the runner.

it dosent consistently failed we have similar(but i not shure if it identical) a couple of month ago and today it returned to behave properly after a week that i had this problem

while debuging i also try to do yaml like
sleep 1000
exit 1
but the line “exit 1” dosent run and i get the same behavior.
i tryed adding resources but it still failed and it didnt look like it used it (according to the k8s metrics)

I self-host gitlab 14.9.3-ee and I use runner version 0.45.0-beta